[GHC] #10534: Data families + Coercible = unsafeCoerce

GHC ghc-devs at haskell.org
Tue Jun 16 01:20:16 UTC 2015


#10534: Data families + Coercible = unsafeCoerce
-------------------------------------+-------------------------------------
              Reporter:  goldfire    |             Owner:
                  Type:  bug         |            Status:  new
              Priority:  highest     |         Milestone:  7.10.2
             Component:  Compiler    |           Version:  7.10.2-rc1
              Keywords:              |  Operating System:  Unknown/Multiple
          Architecture:              |   Type of failure:  None/Unknown
  Unknown/Multiple                   |        Blocked By:
             Test Case:              |   Related Tickets:
              Blocking:              |
Differential Revisions:              |
-------------------------------------+-------------------------------------
 When I say

 {{{
 {-# LANGUAGE TypeFamilies, FlexibleContexts #-}

 module A where

 import Data.Coerce

 data family DF a

 silly :: Coercible (DF a) (DF b) => a -> b
 silly = coerce
 }}}

 {{{
 {-# LANGUAGE TypeFamilies #-}

 module B where

 import A

 newtype instance DF a = MkDF ()

 unsafeCoerce :: a -> b
 unsafeCoerce = silly
 }}}

 I get

 {{{
 [1 of 2] Compiling A                ( A.hs, interpreted )
 [2 of 2] Compiling B                ( B.hs, interpreted )
 Ok, modules loaded: A, B.
 }}}

 Eep!

 Happily, the fix is very very simple: `TyCon.isDistinctAlgRhs` should
 return `False` for `DataFamilyTyCon`s. This fix is already in-flight for
 HEAD in Phab:D968, but this ticket serves to correct this problem in
 7.10.2 before the release.

 Patch coming very shortly.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10534>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list